Key takeaways
- A nonprofit can control its own accounts and data on a volunteer's device, but not the device itself.
- California's customer records law defines a business to include nonprofit organizations.
- Turning off open user consent stops volunteers from granting new AI apps access to organization data with their work account.
- Intune app protection policies work on personal phones without enrolling the device.
- Volunteers who work only from personal email accounts are outside every organization control.
Volunteers bring their own phones and laptops, and AI tools are one tap away on every one of them. A nonprofit cannot block AI on personal devices it does not own, but it can control the two things that matter most: its own sign-in accounts and its own donor and client data. This checklist shows how.
Can a nonprofit stop volunteers from signing up for AI tools?
A nonprofit can stop volunteers from using their organization account to sign up for or connect AI tools, and it can keep organization data from leaving managed apps on personal devices. It cannot stop a volunteer from creating a personal AI account on their own phone. The goal is to keep donor, client and program data inside approved tools.
That distinction shapes everything below. Controls that depend on owning the device, such as blocking websites or inventorying installed apps, do not reach a volunteer's personal phone. Controls tied to the organization's identity and data do.
If you do not yet know which AI apps are connected to your accounts, start with shadow AI detection for nonprofits, then come back to these steps.
Why do AI sign-ups with an organization account put donor data at risk?
An AI tool connected to an organization account can reach the email, files or calendars behind that account, and the nonprofit has no contract covering how the vendor keeps or uses that data.
Donor records, client case notes and program data often include personal information that California law expects organizations to protect. California's customer records law defines a business to include nonprofit organizations, and Civil Code section 1798.81.5 requires businesses that own or license Californians' personal information to maintain reasonable security procedures and practices.
That is a governance gap a board will want closed before a funder or a donor asks about it.
How do you block AI sign-ups that use the organization's Microsoft account?
Turn off open user consent in Microsoft Entra ID so volunteers cannot grant a new app access to organization data on their own. By default, Microsoft 365 lets users consent to apps that request permissions not requiring an administrator. Microsoft's own guidance on user consent recommends allowing consent only for apps from verified publishers, and organizations with many volunteers can go further and require administrator approval for every new app.
- In the Entra admin center, open Enterprise applications, then Consent and permissions, and change User consent for applications.
- Turn on the admin consent workflow so a volunteer who wants a tool can request it instead of hitting a dead end.
- Review apps that already have consent. Changing the setting stops future grants only, and existing grants stay in place until you remove them.
How do you block AI sign-ups that use a Google Workspace account?
Restrict unconfigured third-party apps in the Google Admin console. Under API controls, Google offers a setting that blocks all third-party API access, including Sign in with Google, for any app you have not marked as trusted or limited.
Google also offers a middle option that allows apps requesting only basic sign-in information. That option still lets a volunteer create an account at an AI service with an organization login, so it does not meet the goal on its own. Mark the AI tools you have approved as trusted, and block the rest.
How do you protect organization data on a volunteer's phone?
Use Microsoft Intune app protection policies, which work on personal phones without enrolling the device. Microsoft describes app protection as protecting company data at the app level on devices not enrolled in device management, and the policies apply only when the volunteer is working in the organization's context.
- Restrict cut, copy and paste from Outlook, Teams and OneDrive into personal apps, which covers pasting donor notes into a consumer AI app.
- Control where copies of organization data can be saved.
- Require a PIN to open the work side of the app.
- Selectively wipe organization data when a volunteer leaves, without touching their photos or personal email.
Pair the policies with a Conditional Access rule that requires an app protection policy before email and files open. Microsoft notes that Conditional Access is how you make sure the protection is actually enforced. Conditional Access and Intune are included in Microsoft 365 Business Premium, and Microsoft's Conditional Access planning guide covers licensing and a report-only rollout.
What about volunteers who use personal laptops?
On Windows, a Conditional Access policy can require an app protection policy for browser access to Microsoft 365, which means the volunteer works in Microsoft Edge signed in with the organization account. Microsoft recommends starting in report-only mode and blocking device platforms the policy does not support.
Intune app protection settings for Windows can limit copying, pasting and printing organization data out of that Edge work profile. As of October 2026, Microsoft documents Purview data loss prevention for AI sites in Edge only for Intune-managed Windows devices, so it does not reach a volunteer's personal laptop. A volunteer's personal browser profile is outside these controls.
Why can't you block AI websites on personal devices?
Network and device controls only reach equipment the organization manages. Microsoft Entra tenant restrictions v2, for example, can stop personal Microsoft accounts and outside tenants on the organization's devices and networks, but it does not follow a volunteer home on their own phone. Web filtering works the same way. That is why the steps above focus on accounts and data.
What should a nonprofit do this quarter?
- Write down the AI tools volunteers may use and what data each may touch. Keep it to one page.
- Set a simple approval process for new AI tools so requests from volunteers have somewhere to go.
- Give volunteers who handle donor or client data an organization account, so the controls above can apply to them. Volunteers working only from personal email are outside every control on this list.
- Change user consent in Entra ID or third-party app access in Google Workspace, and turn on an approval request path.
- Review and remove existing app grants that no one approved.
- Deploy app protection policies and a Conditional Access rule in report-only mode, then turn them on after a short test with a few volunteers.
- Add selective wipe to your volunteer offboarding steps.
These steps line up with the CIS Controls safeguards for authorized software and data protection, which gives the board a recognized framework to point to.
If your board has not yet approved an AI policy, a one-page nonprofit AI policy gives these controls a written rule to enforce.
How s90 helps
s90 sets up AI guardrails for nonprofits on the tools they already license, including consent controls, app protection for personal devices and alerts when someone connects a new app to the organization's accounts. We keep the approved tool list current and show you what changed each month. If volunteers are using AI on their own devices today, we are glad to walk through what these controls would look like for your team.
